function f=jFitnessFunction(dataset,X,func_num)

if func_num==1

classif='KNN';

elseif func_num==2

classif='LDA';

elseif func_num==3

classif='RegTree';

elseif func_num==4

classif= 'NB';

elseif func_num==5

classif='SVM';

end

FeatIndexKNN = find(X==1); %Feature Index

if ~any(X(:))

FeatIndexKNN(1) = 1; %Check that at least one feature is selected

end

feat = dataset(:,1:end-1);

data_tr = feat(:,[FeatIndexKNN]);

data_tr=[data_tr dataset(:,end)];

[Ninst, NF] = size(data_tr);

NF = NF - 1;

CVF = 3;

indices = crossvalind('Kfold',Ninst,CVF);

fac = 0.00000000000000000000001;

f = Fit_classifyCV(data_tr(:,1:NF), data_tr(:,end), indices, classif) + fac;

end

function [test_Predicted_labels] = knn(Xtrain,Ltrain,Xtest, K)

%% This is the matlab implemenatation for K nearest neighbors

% Input Arguments:

% Xtrain : training data set

% Ltrain : Labels of training samples

% Xtest : test data set

% K : number of nearest neighbors

% Output Arguments :

% TestLabel : Predicted labels of the output data set

% default value of K if not given by user is 8.

if(nargin < 3)

error('Incorrect number of inputs.');

end

if(nargin < 4)

K = 8;

end

[N , ~] = size(Xtrain);

[Ntest,~] = size(Xtest);

distance = zeros(N,Ntest);

%descendingDistances = zeros(N,Nt);

%Ltest = repmat(Xtest(1,:),N,1);

% calculating the euclidean distance of the test samples from training

% samples

for i = 1: Ntest

for j = 1: N

distance(j,i) = norm(Xtest(i,:)-Xtrain(j,:));

end

end

% ascendingdistances stores all the distances of the test samples

% from the all training samples in cloumns

% Index will have indices of the corresponding training sample

[~,Index]= sort(distance,'ascend');

% consider only top K nearest neighbors to predict the label for test

% sample

Ltest = zeros(K,Ntest);

for i = 1:Ntest

for j=1:K

Ltest(j,i) = Ltrain(Index(j,i));

end

end

test_Predicted_labels = mode(Ltest);

test_Predicted_labels = test_Predicted_labels';

end

function[Error] = Misclassification_error(test_labels, test_Predicted_labels)

%% this is the matlab Implementation for KNN misclassification error

% if you have labels for your test data set

% use KNN.m to find the predicted labels for test samples.

% then you can use this function to find misclassification error for your

% test data set

% Input arguments:

% test_labels : Known labels of your test data

% test_predicted_labels : predicted test labels

if(nargin < 2)

error('Incorrect number of inputs.');

end

% check if the input labels are in row vector

% if yes make it column vectors

if(isrow(test_labels)==1)

test_labels = test_labels';

end

if(isrow(test_Predicted_labels)==1)

test_Predicted_labels = test_Predicted_labels';

end

% matrix dimensions should be equal

[Ltest,~] = size(test_labels);

if (Ltest ~= size(test_Predicted_labels,1))

Error('matrix dimensions are not consistent');

end

count=0;

for i=1:Ltest

if(test_labels(i,1)==test_Predicted_labels(i,1))

count = count+1;

end

end

missPredictions = count;

Error = missPredictions / Ltest;

end

function Fit = Fit_classifyCV(data, dataLab, indices, classif)

CVF = max(indices);

Fit = zeros(CVF,1);

for k=1:CVF

testn = (indices == k);

trainn = ~testn;

%NTest = sum(testn);

switch classif

case 'KNN'

test_Predicted_labels = knn(data(trainn,:),dataLab(trainn),data(testn,:),3);

Fit(k) = Misclassification_error(dataLab(testn),test_Predicted_labels);

end

end

Fit =1 - mean(Fit);

end